Yes, female premiums went up to cover the cost of not having that factor in the model.
Fortunately I'm sure people are hard at work finding "proxy factors" which happen to map to pretty well gender but have an acceptable parallel explanation for why they're meaningfully descriptive of an applicants risk profile :)